An Immersive Dive into the Definition of Science Literacy

Introduction 

In the contemporary digital age, science literacy has emerged as a crucial form of competency for individuals to navigate and understand the complexities of the modern world. Traditionally, science literacy was defined as the ability to possess an adequate understanding of scientific concepts necessary for making informed decisions, collective economic benefit, and participation in other organizational responsibilities (National Research Council, 1996). Recent developments including changes to educational and healthcare institutions, have expanded this definition to include the ability to evaluate and interpret scientific information, critically assess the credibility of sources, and understand the ethical implications of scientific advancements through a socioeconomic lens (Feinstein, 2011). As society rapidly increases its dependence on science and technology, the importance of science literacy has become imperative.

The advent of social media has significantly influenced the landscape of science literacy. Platforms such as X (formerly Twitter), Facebook, Instagram, TikTok, and YouTube have democratized access to scientific information, allowing for the rapid dissemination and discussion of scientific ideas among a broad audience. Social media has enabled scientists to share their research directly with the public, engage in real-time discussions, and collaborate across different geographic locations (Pew Research Center, 2018). For instance, a survey conducted by the Pew Research Center (2018) found that 36% of Americans use social media, among other platforms such as magazines and documentaries to access scientific news, highlighting the increasing reliance on these platforms for scientific information (Funk et al., 2017). This shift presents both opportunities and challenges for enhancing science literacy in the general public which will be discussed in the following sections.

The COVID-19 Pandemic, Social Media, and the Spread of Harmful Information 

The COVID-19 pandemic exemplifies how social media can amplify misinformation, leading to what the World Health Organization has termed an “infodemic”—an overwhelming flood of information, both accurate and inaccurate, that makes it difficult for individuals to find reliable sources and make informed decisions (Cinelli et al., 2020). During the pandemic, misinformation about the virus’ origin, transmission, prevention, and treatment spread rapidly on social media, complicating public health efforts and contributing to widespread confusion and fear (Cinelli et al., 2020). For example, a study published in the Journal of Medical Internet Research found that over one-quarter of the most viewed YouTube videos on COVID-19 contained misleading information (Li et al., 2020). Specific instances of misinformation identified in other research included false claims about COVID-19 cures, such as drinking bleach or taking hydroxychloroquine without medical supervision (Gharpure et al., 2020). 

Disinformation campaigns, often driven by malicious individuals or groups, spread conspiracy theories about the origins of the virus, including the false narrative that it was engineered as a bioweapon (Kucharski & Russell, 2020). Fear-mongering also played a significant role, with exaggerated reports of the virus’ effects leading to risky and harmful behaviours, such as hoarding essential supplies, inadvertently contributing to the spread of the disease (Pennycook et al., 2020). Furthermore, misinformation about the effectiveness of lockdown protocols, vaccines, and masks further complicated public compliance and trust in health authorities. For instance, a study conducted by Brennen et al. (2020) found that nearly 59% of the false claims (claims that are scientifically incorrect) about COVID-19 were rated as “false” (simply incorrect), while approximately 40% of those claims were labeled “misleading,” contributing to public distrust and undermining health efforts.

Current Gaps in Knowledge

This regular interaction between the general public and scientific information on social media underscores the importance of science literacy in the digital age. It is imperative for individuals to develop the skills necessary to discern credible scientific information from misinformation or disinformation. Understanding how social media influences the public’s perception of science is critical for developing strategies to enhance science literacy and mitigate the spread of misinformation. For instance, organizations such as the “COVID-19 Infodemic Observatory,” which monitors and analyzes misinformation on social media, have been pivotal in identifying and correcting false information (Gallotti et al., 2020). Initiatives like these exist globally and are vital for society to develop and foster trust so that misinformation can be combated.

Understanding how perceptions, emotions, attitudes, and comprehension of scientific information are shaped through social media is critical, as the COVID-19 pandemic has shown how misinformation can influence public health outcomes and social behaviour. This highlights the importance of the perception of scientific information as it can directly affect the willingness to engage with the content and make informed decisions (Fischhoff & Scheufele, 2013). Consequently, research into how different demographics perceive and interact with scientific information online can inform targeted educational interventions and policies to improve science literacy. However, there is a notable lack of research in this area, representing a significant knowledge gap that this manuscript aims to address.

This research study explores the role of social media in shaping the public’s understanding of science by examining the different perceptions, attitudes, and overall comprehension of the general public of COVID-19-related information on X, formerly known as Twitter. Researchers explored the varying reactions to content and evaluated the original content itself, within the time frame of March 2020 to May 2023 (the time period of the public emergency state of the pandemic in Canada). It is expected that there will be variation in the perceptions and opinions regarding COVID-19-related content, specifically in the overall understanding of COVID-19-related content and the frequency of certain emotions. Through in-depth analysis, this will shine light on the current state of science literacy among the general public.

Methodology 

Protocol Development

The protocol development consisted of the following factors: the social media platform(s) examined, the specific keywords searched for within posts, the number of key posts, and the time frame. 

After examining the methodologies of past research studies evaluating social media content, researchers decided to utilize Google Trends to determine the four most prevalent search terms on social media platforms from the March 2020 to May 2023 time frame (Cinelli et al., 2020). These terms were found to be: “Vaccination,” “COVID Vaccine,” “Coronavirus,” and “Vaccine.” 

When exploring different social media platforms to evaluate, Instagram, Facebook, and X (formerly Twitter) were found to have the greatest engagement with content pertaining to the COVID-19 pandemic (Pennycook et al., 2020). However, both Instagram and Facebook were found to have restrictions on complete access to dated posts (COVID-19-related content were blocked due to pop-ups and the information was often blurred), preventing researchers from drawing comparisons across different platforms. Further, the user profiles for the studied posts were inaccessible. As a result, X (formerly Twitter) was the sole social media platform used for this study.

Data Collection and Scoring

The four search terms found via Google Trends’ COVID-19-related queries were used to filter tweets found on X (formerly Twitter) from the aforementioned time period. These search terms were inputted in X’s advanced filter function (along with the time period). Researchers gathered the top ten most interacted with tweets from each keyword, as well as the top ten replies from each keyword, creating a total of n = 40 tweets and n = 400 replies. The top ten most interacted with tweets were chosen because posts beyond this point lacked sufficient engagement for analysis. Each reply and tweet was stored on a Google Sheet, along with supplemental information such as a link to the tweet, the type of content shared (media), the poster’s handle, and the number of total interactions at the time (likes, retweets, and replies). After data collection was completed, each researcher scored every tweet and reply across all four keywords independently to simulate a blinding measure to minimize experimenter bias. The blinding method entailed each researcher scoring their replies and tweets on a separate sheet, until all the data was complete, upon which all the data was collected and integrated on a single sheet for analysis. The resulting data was then summarized in a new table used for statistical analysis.

Rubric

In order to evaluate the different aspects of the content of the original post, a rubric was created. Table 1 (Appendix) was used to evaluate different components of the original tweet. These included content relevance, tone, language and clarity, context awareness, sarcasm, inclusion of negative words, and the overall understanding of scientific content displayed by the poster. With the exception of the overall understanding score, these metrics were not used in the final analysis but will serve as data for more specific, future research. These categories were scored on a Likert scale (1-5) with a higher number corresponding to greater prevalence of that category. Likert scales were also used to reduce biases present from the different perspectives of the researchers. 

Researchers assessed the replies to each original tweet using the following categories: rejection, anger/frustration, acceptance/calmness, and confusion. Each reply was ranked as belonging to one of the four categories by a corresponding number. Replies affiliated with rejection were labeled 0-2, anger/frustration were labeled 3-5, confusion was labeled 6-8, and acceptance/calmness were labeled 9-10. Replies reflecting sarcasm, used a non-English language, and/or were irrelevant to the scope of the study, were omitted from the ranking and scored as N/A (a total of 21 replies, producing n = 379 which were used for analysis after omissions). Ultimately, the frequency of each emotion across the replies of all tweets was recorded for analysis.

Statistical Analysis

Four chi-square tests were conducted using Google Sheets to test for the presence of a statistically significant difference using a non-parametric analysis. The tests were done to assess a difference in the prevalence of a specific emotion (rejection, anger/frustration, confusion, and acceptance/calmness) out of all four emotions in a single keyword (“Vaccination,” “COVID Vaccine,” “Coronavirus,” and “Vaccine”). This was carried out for all four keywords. A follow-up z-test was performed to make pairwise comparisons among the four groups with a 95% confidence interval.

To supplement the chi-square tests, a single factor ANOVA using Google Sheets was used to identify differences in overall understanding of the main tweets across the four keywords. Consequently, this test was performed for each of the search terms and for the overall understanding scores of the original tweets. To compare the results, a two-sample t-test assuming equal variance was performed between each term, and any statistically significant results were highlighted.

Results 

Overall Understanding Across Different Keywords

A single-factor ANOVA was performed to identify differences in overall understanding ratings across the four keywords (“COVID Vaccine,” “Coronavirus,” “Vaccination,” and “Vaccine”). The omnibus test from ANOVA indicated a statistically significant difference in the overall understanding across these keywords (F(3, 116)=2.93, p = 0.036). This F-statistic implies that there is a significant difference in the average understanding ratings between at least two of the keyword groups. Follow-up t-tests were performed to make pairwise comparisons, corrected using Bonferroni correction. Results indicated that overall understanding ratings in the keyword “Vaccine” were significantly higher than in the “COVID Vaccine” keyword (Figure 1; t(58) = -3.58, p = 0.001). This finding provides evidence that on social media, there may be a better overall understanding of vaccines in general, as opposed to COVID vaccines specifically.